This repository has been archived by the owner on Jul 16, 2024. It is now read-only.
Support more index types besides ivfflat #224
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
beeender
reviewed
Nov 3, 2023
I've proposed PR #225 for installing pgvector on CI. After merging that PR, we can add test cases for embedding. |
yihong0618
reviewed
Nov 14, 2023
Comment on lines
+119
to
+125
| import sentence_transformers # type: ignore reportMissingImports | ||
|
|
||
| model = sentence_transformers.SentenceTransformer(model_name) # type: ignore reportUnknownVariableType | ||
| embedding_dimension: int = model[1].word_embedding_dimension # From models.Pooling | ||
| except: | ||
| raise NotImplementedError( | ||
| "Model '{model_name}' doesn't provide embedding dimension information" |
There was a problem hiding this comment.
maybe import error here
There was a problem hiding this comment.
I think import error is for modules, not models.
| ) | ||
| if method is not None: | ||
| assert method == "hnsw" or method == "ivfflat" |
There was a problem hiding this comment.
since there may be more method , assert method in ["ivfflat", "hnsw"] may be bettor?
There was a problem hiding this comment.
Thanks! That's easier to read and write indeed. Changed.
yihong0618
approved these changes
Nov 14, 2023
added 17 commits
November 18, 2023 01:45
|
A new test case for embeddings is added in To support |
beeender
approved these changes
Nov 23, 2023
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Previously, we only support indexing embeddings using the
ivfflataccess method in pgvector.
Recently, a new access method
hnswhas been added to pgvector.hnswis believed to be more performant and accurate thanivfflat.To allow for more flexibility, we add a new parameter
methodtoallow user to choose which access method to use when creating index.
Also, a new parameter
embedding_dimensionis added to support moremodels, since dimension is required for pgvector to create index.
A new test case for embeddings is added in
tests/test_embedding.py.To support
set allow_system_table_mods = on;, Postgres is upgradedfrom 12 to 13 on CI.